vRealize Automation 7 – Load Balancer RulesFebruary 24, 2016
In a previous post we went over installing an Enterprise Install of vRealize Automation behind a load balancer. This install required us to setup a Load Balancer with three VIPs but also required that we only had one active member in each VIP. A load balancer with a single member doesn’t really balance much load does it?
After the installation is done, some modifications need to be made on the Load Balancer. The instructions on this can be found in the official vRealize Automation Load Balancing Configuration Guide if you want to learn more. There are several examples on how to setup load balancing on an F5 load balancer and NSX for example. This post will focus on a KEMP load balancer which is free for vExperts and it will all be shown through with GUI examples.
IaaS Manager Service
Let’s start with the Iaas Manager Service. Modify your IaaSmgmt service VIP and make sure the load balancing method is “Round Robin”. Then we want to add some health check parameters. the URL that we should check is located at /VMPSProvision and we’re going to run a “GET” command on this. The return response should be “ProvisionService” without the quotes.
Now we’ll move onto the web services VIP. Let’s modify our IaaSweb VIP and start with updating the persistence options. We want session persistence to be based on the Source Address and have a timeout of 30 minutes. The Load Balancing policy should again be Round Robin and now we can move on to the health check.
Remember at the end of the Enterprise Install, we’re given a screen with Load Balancer Information that we should go back and check? Well, the Web Services health check has the wrong case. Notice that the Monitor for the IaaS Web component says “registered” but it should actually be “REGISTERED.”
If we check the URL we can see the monitor clearly.
Finish adding your health check with a URL of /wapi/api/status/web with a GET method that is looking for a pattern named “REGISTERED.”
Now we can do the vRealize Appliance URL. Edit the VIP and add an additional port of 8444 to the existing 443. Port 8444 is used for remote console access which is a useful access method that you might want. Change the persistence options to a source based method, with a timeout of 30 minutes just like we did for the web services. The load balancing method is going to be “Round Robin” like it has been for our other services, and then it’s time to do our health check again.
This time we want to look for a URL of /vcac/services/api/health with a GET method, and we’re only looking for a 200 or 204 response pattern back so nothing needs to be added in the “Reply Pattern” box.
Now we’ve added all of our load balancing rules and we can see our VIPs are all up and happy with our members. The enterprise environment now has some failover capabilities that a simple installation is lacking.